Project Summary/Abstract: The nervous system must constantly compare different streams of information, make decisions, and execute appropriate behaviors. However, the neural mechanisms of decision-making are not yet fully understood. This project uses Drosophila as a model to investigate sensorimotor integration and decision-making, and in particular the origins of variability in decision outcomes. Fruit flies are a useful model for studying these issues because there are excellent genetic tools in Drosophila that provide access to many individual cell types, and it is straightforward to monitor neural activity in an awake behaving organism. The number of neurons in each brain region is relatively small (on the order of 100 ? 1000), and the wiring of the brain is relatively stereotyped. Finally, almost the entire brain can be imaged in a single experiment, due to the relatively small size of the brain. This project will focus on a well-characterized behavior called osmotropotaxis, where a fly compares odor concentrations at the left and right antenna. On average, a fly will turn in the direction of the more strongly stimulated antenna. On a given trial, however, the fly may turn in the opposite direction or simply walk straight ahead, even when olfactory signals are unambiguous at the level of the antennae, indicating that behavioral variability arises internally, and is not the result of sensory noise. This project has two Specific Aims: (1) to determine how the fly compares the level of activity on the right and left sides of its olfactory system, and (2) to determine how the outcome of this computation interacts with internal dynamics in the brain to control behavioral decisions. These aims will be accomplished using volumetric two-photon calcium imaging and whole-cell patch clamp recordings in tethered flies executing osmotropotaxis behavior on a spherical treadmill. Neural activity will be monitored in several specific brain regions near the interface of the olfactory system and motor control systems. The feasibility of this approach has already been demonstrated in pilot experiments. The first part of the project aims to identify and characterize neurons that explicitly encode a comparison between the right and left antenna ? in other words, neurons that respond preferentially to laterally asymmetric odor stimuli. The second part of the project aims to identify and characterize neurons whose activity can predict trials where the fly fails to turn in the cued direction. Our overall goal is to understand how the brain explicitly ?decides? whether the odor is on the right or left side, and how a clearly asymmetric stimulus is translated into an unreliable turning behavior. We expect this project to provide insights into the fundamental features of sensorimotor integration, decision-making, and behavioral variability.